Beginner perfumer · 106
The median of 954 public formulas is 17 materials, and the largest component is a quarter of the whole
· 13 min read
'Commercial perfumes contain hundreds of ingredients' is a thing people say. I have 954 deduplicated public demonstration formulas, and their median size is 17 materials, with nearly half (47.5%) falling between 10 and 19, only 11.0% above 30, and just four above 50. The distribution of component shares is equally clear: across all 17,481 components the median share is 2.50%, four in ten fall between 1% and 5%, and only 6.0% exceed 20%. The largest single component in a formula is a median of 25.0% of the total. The most practical number is the ratio: the largest component divided by the smallest has a median of 60 — which determines what scale you need, and why you make dilutions.
About a 5 minute read.
I have used this corpus for name matching, but never asked a more basic question: what does a formula look like?
A median of 17
Material counts across the 954 formulas:
| Materials | Formulas | Share |
|---|---|---|
| 5–9 | 143 | 15.0% |
| 10–19 | 453 | 47.5% |
| 20–29 | 253 | 26.5% |
| 30–49 | 101 | 10.6% |
| 50+ | 4 | 0.4% |
Median 17, mean 18.3, maximum 57.
Nearly half sit between 10 and 19 materials.
"Commercial perfumes contain hundreds of ingredients" does not hold in this corpus. (These are demonstration formulas published by fragrance houses, not reverse-engineered retail perfumes — a limit I return to at the end.)
Component shares
The distribution across all 17,481 component entries:
| Share | Entries | Proportion |
|---|---|---|
| < 0.1% | 243 | 1.4% |
| 0.1–1% | 4,120 | 23.6% |
| 1–5% | 7,002 | 40.1% |
| 5–20% | 5,061 | 29.0% |
| ≥ 20% | 1,055 | 6.0% |
Median 2.50%. Four in ten components sit between 1% and 5%.
And within a formula, the share of components below 1% has a median of 15.1% — in a 17-material formula, two or three are trace.
The biggest one is a quarter
The largest single component's share of the total: median 25.0%, mean 28.8%, and 7.5% of formulas have one material above 50%.
A typical formula has one clear body and a crowd of modifiers.
And the most practical number is the ratio
Largest component ÷ smallest:
| Median | 60× |
| Mean | 214× |
| Maximum | 9,333× |
| Formulas above 1000× | 28 (2.9%) |
This number decides what equipment you need.
Say you are making 10 grams. The largest component at 25% is 2.5 grams. A ratio of 60 puts the smallest at 42 milligrams.
42 milligrams is not something most household scales weigh accurately. And 2.9% of formulas span three orders of magnitude — that is not a scale problem, it is the reason dilutions exist.
This is why every introductory book tells you to make 10% and 1% dilutions first. Not only to protect your nose (though it does that too), but because a formula's own dynamic range exceeds a scale's resolution.
The most frequently used materials
And the corpus hands over a real starter kit — ranked by how many formulas contain them:
| Formulas | Share | Material |
|---|---|---|
| 445 | 46.6% | dipg (dipropylene glycol, a solvent) |
| 338 | 35.4% | Linalool |
| 317 | 33.2% | Benzyl acetate |
| 274 | 28.7% | Phenethyl alcohol |
| 218 | 22.9% | Citronellol |
| 218 | 22.9% | Coumarin |
| 207 | 21.7% | Linalyl acetate |
| 192 | 20.1% | Geraniol |
| 165 | 17.3% | Hedione |
| 160 | 16.8% | Hydroxycitronellal |
| 159 | 16.7% | Benzyl salicylate |
| 157 | 16.5% | Eugenol |
| 157 | 16.5% | α-hexyl cinnamaldehyde |
| 150 | 15.7% | Lilial |
| 139 | 14.6% | Heliotropin |
| 139 | 14.6% | Dihydromyrcenol |
| 132 | 13.8% | Patchouli oil |
| 131 | 13.7% | Bergamot oil |
All eighteen already have their own article in this series. (That was not planned — I picked them one at a time for other reasons.)
And this list is worth setting against article 95's failure. There I tried to compute a starter kit from the official blending suggestions and got an echo chamber of sixteen caramellic materials.
This time no computing is needed. Counting is enough. And the count is entirely sensible: one solvent, four basic floral alcohols and esters, coumarin, Hedione, a salicylate and an aldehyde, patchouli and bergamot.
For "which materials get used", the formula corpus answers far better than the blending network does.
In practice
- Seventeen is a reasonable target. Do not assume a formula has to be complicated to count.
- Decide the quarter-share body first. The rest is modification.
- Make 10% and 1% dilutions, because the ratio is 60. Not fastidiousness — arithmetic.
- Those eighteen are a shopping list. Not anybody's opinion; a count across 954 formulas.
What this doesn't establish
- These are fragrance houses' published demonstration formulas, made to show how certain materials are used — not the actual formulas of retail perfumes. Real commercial formulas may be more complex or less; I have no data.
- Demonstration formulas skew systematically toward the materials being demonstrated, so that frequency list is biased.
- Shares are computed from the listed figures, and some formulas omit solvent entirely (87.4% list none), so those percentages are shares among the listed components.
- Material names are not normalised, so different spellings of one material (a problem I measured) are counted separately — the true frequencies may be higher.
- "42 milligrams is unweighable" depends on your scale. That is illustrative arithmetic, not a specification.
- I have not made any of these formulas.